Yes you can. Anomaly Detectors can handle missing values in your dataset. Missing values are handled in the same way as the BigML models and ensembles. If you would like to know how BigML handles missing values to calculate predictions with your models and ensembles, please read this question.
